Find Command In Unix With Examples

Shell Scripting

Search based on file name or regular expression match

The -name argument specifies a matching string for the filename. We can pass wildcards as its argument text. *.txt matches all the filenames ending with .txt and prints them. The –print option prints the filenames or file paths in the terminal that matches the conditions (for example, –name) given as options to the find command.

$ find /home/slynux -name “*.txt” –print

The find command has an option –iname (ignore case), which is similar to -name. –iname matches the name ignoring the case.

For example:

$ ls

example.txt EXAMPLE.txt file.txt

$ find . -iname “example*” -print

./example.txt

./EXAMPLE.txt

If we want to match either of the multiple criterions, we can use OR conditions as shown below:

$ ls

new.txt some.jpg text.pdf

$ find . \( -name “*.txt” -o -name “*.pdf” \) -print

./text.pdf

./new.txt

The previous code will print all of the .txt and .pdf files, since the find command matches both .txt and .pdf files. \( and \) is used to treat -name “*.txt” -o –name “*.pdf” as a single unit.

The -path argument can be used to match the file path for files that match the wildcards. -name always matches using the given filename. However, -path matches the file path as a whole.

For example:

$ find /home/users -path “*slynux*” -print

This will match files as following paths.

/home/users/list/slynux.txt

/home/users/slynux/eg.css

The -regex argument is similar to -path, but -regex matches the file paths based on regular expressions. Regular expressions are an advanced form of wildcard matching. It enables to specify a text

with patterns. By using the patterns, we can make matches to the text and print them. A typical example of text matching using regular expressions is: parsing all e-mail addresses from a given pool of text. An e-mail address takes the form [email protected]. So, it can be generalized as [a-z0-9]+@[a-z0-9]+.[a-z0-9]+. The + signifies that the previous class of characters can occur one or more times, repeatedly, in the characters that follow.

The following command matches .py or .sh files:

$ ls

new.PY next.jpg test.py

$ find . -regex “.*\(\.py\|\.sh\)$”

./test.py

Similarly, using -iregex ignores the case for the regular expressions that are available.

For example:

$ find . -iregex “.*\(\.py\|\.sh\)$”

./test.py

./new.PY

Negating arguments find can also take negation of arguments using “!”. For example:

$ find . ! -name “*.txt” -print

The above find construct matches all the file names, as long as the name does not end with .txt. The following example shows the result of the command:

$ ls

list.txt new.PY new.txt next.jpg test.py

$ find . ! -name “*.txt” -print

.

./next.jpg

./test.py

./new.PY

Search based on the directory depth

When the find command is used it recursively walks through all the subdirectories as much as possible until it reaches the leaf of the subdirectory tree. We can restrict the depth to which the find command traverses using some depth parameters given to the find. –maxdepth and -mindepth are the parameters.

For specifying max depth we use the –maxdepth level parameter. Similarly, we can also specify the minimum level at which the descending should start. If we want to start searching from the second level onwards, we can set the minimum depth using the –mindepth level parameter. Restrict the find command to descend to a maximum depth of 1, by using the following command:

$ find . –maxdepth 1 –typr f –print

This command lists all the regular files only from the current directory. If there are subdirectories, they are not printed or traversed. Similarly, -maxdepth 2 traverses up to at most two descending levels of subdirectories.

-mindepth is similar to –maxdepth, but it sets the least depth level for the find traversal. It can be used to find and print the files that are located with a minimum level of depth from the base path. For example, to print all the files that are at least two subdirectories distant from the ccurrent directory use the following command:

$ find . –mindepth 2 –type f –print

./dir1/dir2/file1

./dir3/dir4/f2

Even if there are files in the current directory or dir1 and dir3, it will not be printed.

Search based on file type

The file search can be filtered out using the –type option. By using –type, we can specify to the find command that it should only match files having a specified type.

List only directories including descendants as follows:

$ find . -type d -print

It is hard to list directories and files separately. But find helps to do it. List only regular files as follows:

$ find . -type f -print

List only symbolic links as follows:

$ find . -type l -print

You can use the type arguments from the following table to properly match the required file type:

File types and there corresponding arguments

 

Search on up file times

UNIX/Linux file systems have three types of timestamp on each file. They are as follows:

  • Access time (-atime): This is the last timestamp of when the file was accessed by some user
  • Modification time (-mtime): This is the last timestamp of when the file content was modified
  • Change time (-ctime): This is the last timestamp of when the metadata for a file (such as permissions or ownership) was modified

 

There is nothing called creation time in UNIX.

-atime, -mtime, -ctime are the time parameter options available with find. They can be specified with integer values in “number of days”. These integer values are often attached with – or + signs. The – sign implies less than whereas the + implies greater than.

For example:

  • Print all the files that were accessed within the last 7 days as follows:

$ find . -type f -atime -7 –print

 

  • Print all the files that are having access time exactly 7 days old as follows:

$ find . -type f -atime 7 -print

  • Print all the files that are having access time older than 7 days as follows:

$ find . -type f -atime +7 -print

Similarly, we can use the –mtime parameter for search files based on modification time and -ctime for search based on change time.

-atime, -mtime, and –ctime are time-based parameters that use the time metric in days. There are some other time-based parameters that use the time metric in minutes. These are as follows:


  • -amin (access time)
  • -mmin (modification time)
  • -cmin (change time)

For example:

In order to print all the files that are having access time older than seven minutes, use the following command:

$ find . -type f -amin +7 -print

Another nice feature available with find is the –newer parameter. By using -newer, we can specify a reference file to compare with the timestamp. We can find all the files that are newer (older modification time) than the specified file with the –newer parameter.

For example, find all the files that are having a modification time greater than that of the modification time of a given file.txt file as follows:

$ find . -type f -newer file.txt -print

Timestamp manipulation flags for the find command are very useful for writing system backup and maintenance scripts.

Search based on file size

Based on the file sizes of the files, a search can be performed as follows:

$ find . -type f -size +2k

# Files having size greater than 2 kilobytes

$ find . -type f -size -2k

# Files having size less than 2 kilobytes

$ find . -type f -size 2k

# Files having size 2 kilobytes

Instead of k we can use deferent size units as the following:

  • b – 512 byte blocks
  • c – bytes
  • w – two byte words
  • k – Kilobyte
  • M – Megabyte
  • G – Gigabyte

Deleting based on the file matches

The -delete flag can be used to remove files that are matched by find.

Remove all the .swp files from the current directory as follows:

$ find . -type f -name “*.swp” -delete

Match based on the file permissions and ownership It is possible to match files based on the file permissions. We can list out the files having

specified file permission as follows:

$ find . -type f -perm 644 -print

# Print files having permission 644

As an example usage case, we can consider the case of Apache web server. The PHP files in the web server require proper permissions to execute. We can find out the PHP files that are not having proper execute permissions as follows:

$ find . –type f –name “*.php” ! -perm 644 –print

We can also search files based on ownership of the files. The files owned by a specific user can be found out using the -user USER option. The USER argument can be a username or UID.

For example, to print the list of all files owned by the user slynux, you can use the following command:

$ find . -type f -user slynux -print

Executing commands or actions with find

The find command can be coupled with many of the other commands using the –exec option. -exec is one of the most powerful features that comes with find. Let’s see how to use the –exec option.

Consider the example in the previous section. We used –perm to find out the files that do not have proper permissions. Similarly, in case where we need to change the ownership of all files owned by a certain user(Ex: root) to another user. We can find all the files owned by root by using the –user option and using –exec to perform ownership change operation.

Example:

# find . –type f –user root –exec chown slynux {} \;

In this command {} is a special string used with the –execc option. For each file match {} will be replaced with the file name in place for –exec . For example, if the find command finds two files test1.txt and test2.txt with owner slynux, the find command will perform

chown slynux {}

This gets resolved to chown slynux test1.txt and chown slynux test2.txt

Another usage example is to concatenate all the C program files in a given directory and write it to a single file all_c_files.txt. We can use find to match all the C files recursively and use the cat command with –exec flag as follows:

$ find . –type f –name “*.c” –exec cat {} \; > all_c_files.txt

-exec is followed with any command. {} is a match. For every matched filename, {} is replaced with filename. To redirect the data from find to the all_c_files.txt file, we used the > operator instead of >> (append) because the entire output from the find command is a single data stream(stdin) >> is necessary only when multiple data streams are to be appended to a single file.

For example, to copy all the .txt files that are older than 10days to a directory OLD, use the following command:

$ find . –type f –mtime +10 –name “*.txt” –exec cp –f {} OLD \;

-exec can be coupled with printf to produce a very useful output. For example:

$ find . -type f -name “*.txt” -exec printf “Text file: %s\n” {} \;

 

Skip specified directories from the find

Skipping certain subdirectories for a performance improvement is sometimes required while doing a directory search and performing some action. For example, when programmers look for particular files on a development source tree, which is under a version control system such as Git, the source hierarchy will always contain the .git directory in each of the subdirectories (.git stores version control related information for every directory). Since version control related directories do not produce useful output, they should be excluded from the search. The technique of excluding files and directories from the search is known as pruning. It can be performed as follows:

$ find devel/source_path \( -name “.git” -prune \) -o \( -type f -print \)

# Instead of \( -type -print \), use required filter.

The above command prints the name (path) of all the files that are not from the .git directories. Here, \( -name “.git” -prune \) is the exclude portion, which specifies that the .git

directory should be excluded and \( -type f -print \) specifies the action to be performed. The actions to be performed are placed in the second block -type f –print (the action specified here is to print the names and path of all the files).

In case of any ©Copyright or missing credits issue please check CopyRights page for faster resolutions.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.